How long does it take to form a habit? (Not 21 days)
The real numbers from the studies that tracked people building new habits, where the 21-day figure came from, and a plan that expects a few bad weeks.
A new daily habit takes about two months to feel automatic for a typical person, and much longer for many. In the best-known real-world study, the estimates ranged from 18 to 254 days. A 2024 review of health-habit studies puts that study’s median at 66 days and another study’s at 59, with even wider differences between people.
In short
- About two months in the middle. A 2024 review of health-habit studies found medians of 59 to 66 days, including 66 days for the best-known study, from 2010.
- The spread is enormous. Individual times ran from 18 to 254 days in the 2010 study and from 4 to 335 days across the review.
- One missed day made little difference. In the 2010 study, missing a single day did not materially slow habit formation.
- Repetition in the same situation is what counted. In both real-world studies, doing the behaviour consistently after the same cue went with habits forming.
Your turn: Write your new habit as “After I ___, I will ___”, and add the smallest version you would still do on a bad day.
On this page
Where does “21 days” come from?
The 21-day figure is usually traced to a 1960 self-help book and its anecdotes about surgery patients. A 2012 article in the British Journal of General Practice, co-written by the lead author of the main habit study, puts it this way:
“Some patients may have heard that habits take 21 days to form. This myth appears to have originated from anecdotal evidence of patients who had received plastic surgery treatment and typically adjusted psychologically to their new appearance within 21 days.”
Source 2, p. 665
The book they cite is Psycho-Cybernetics by Maxwell Maltz.2 Getting used to a new face is a different process from making a new behaviour automatic. The same authors suggest telling people to expect a new daily habit to take “around 10 weeks.”2
What did the real-world study find?
In a 2010 study, a new daily habit took from 18 to 254 days to get close to its peak automaticity, and a 2024 review gives the median as 66 days. Phillippa Lally and colleagues asked volunteers to pick one eating, drinking or activity behaviour and do it daily in the same situation, such as “after breakfast”, for 12 weeks.1
Every day, people rated how automatic the behaviour felt and whether they had done it. Researchers then fitted a curve to each person’s ratings. Automaticity rose quickly at first and then levelled off. The “days to form a habit” is the point where each person’s curve reached 95% of its plateau.1
The numbers come with caveats. The study ran for 84 days, so any figure above 84 is where a person’s curve was heading, projected forward.1 A 2012 article by two of the study’s authors gives 66 days as an average, and the 2024 review gives it as the median.23
What the research says
ObservationalFor simple daily health behaviours, habits took weeks to months to become automatic, and the time varied hugely from person to person.1
What this can’t tell us: How long a complex habit takes, such as learning an instrument. The behaviours were simple, people chose their own, automaticity was self-rated, and fewer than half of the volunteers fitted the curve well.
What does the 2024 review add?
A 2024 systematic review found the same rough middle, about two months, and an even wider spread. Ben Singh and colleagues at the University of South Australia gathered 20 studies.3 The behaviours included physical activity, healthy eating, drinking water, flossing and taking vitamins.
Only four of the 20 studies measured how long a habit took to form. Their medians ranged from 59 to 66 days, with individuals ranging from 4 to 335 days.3 The reviewers also rated most of the studies at high risk of bias.
| Lally et al., 20101 | Keller et al., 20214 | Singh et al., 20243 | |
|---|---|---|---|
| Design | Daily tracking, 12 weeks | Randomised trial, 84 days | Systematic review and meta-analysis |
| Who | 96 volunteers | 192 adults aged 18–77 | 20 studies, 2,601 participants |
| Behaviour | One eating, drinking or activity habit | One everyday nutrition habit | Health behaviours: activity, diet, water, flossing and others |
| Time to habit | Range 18–254 days; median 66, as the 2024 review reports it | Median 59 days, for those who formed a habit | Medians 59–66 days, means 106–154, range 4–335 |
| Also found | Missing one day did not materially affect it | A cue tied to a routine and a cue tied to a clock time worked equally well | Morning practice and self-chosen habits showed greater strength |
Does missing a day reset the clock?
No. In the 2010 study, a single missed day did not materially slow habit formation. The paper’s abstract states it plainly:
“Missing one opportunity to perform the behaviour did not materially affect the habit formation process.”
Source 1, Abstract
The 2012 summary by the same group adds that “automaticity gains soon resumed after one missed performance.”2 What the study looked at was single misses. It does not tell us what happens after a week away. Doing the behaviour more consistently did go with a curve that fitted better, so repetition still matters.1
What makes a habit form faster?
The most consistent finding is repetition in the same situation, and a few other factors may help.
- Same cue, every day. In a 2021 trial, adults linked a nutrition habit either to a routine (“after breakfast”) or to a clock time (“at 9:00”). Both worked equally well, and repeating the plan was a key predictor of automaticity.4
- Simple actions. In the 2010 study, automaticity peaked sooner for simple actions, such as drinking water, than for more elaborate ones, such as doing 50 sit-ups.2
- Mornings, maybe. The 2024 review found that morning practice and self-selected habits generally showed greater strength. One included study found stretching became a habit faster in the morning than in the evening.3
- Your own choice. People who picked their own habit tended to build stronger ones in the review’s data.3
What this can’t tell us. Most of these findings come from single studies of simple health behaviours. They are worth trying, and none of them is a guarantee.
How do you plan a habit for weeks that go wrong?
Plan for about ten weeks, tie the habit to one daily cue, and decide in advance what the smallest acceptable version is. This plan follows the research above. The 10-week horizon and the weekly rating come from the 2012 article’s advice for patients.2
- Choose one cue you meet every day. “After I put the kettle on” and “at 7:30” both count. The 2021 trial found no difference between them.
- Write a full version and a minimum version. The full version is what you do on a normal day. The minimum version is the smallest action that still happens at the same cue.
- Put a review date ten weeks out. Until then, hold off on judging how it’s going.
- Mark each day yes or no. A missed day is one empty box. The next day, go back to the same cue.
- Once a week, rate how automatic it felt from 1 to 5. Watching it creep up shows progress that a simple count of days can miss.
| Your plan | Example | |
|---|---|---|
| Cue: after I… or at… | After I put the kettle on | |
| Full version | Ten minutes of stretching | |
| Minimum version | One stretch, held for 30 seconds | |
| Review date (10 weeks from today) | ||
| How automatic, weeks 1–5 (1–5 each) | 1 · 2 · 2 · 3 · 3 | |
| How automatic, weeks 6–10 (1–5 each) |
For a whole daily routine built the same way, see a routine that survives bad days.
Questions people ask
Is it 21 days or 66 days to form a habit?
Neither number is a rule. The 21-day figure comes from anecdotes about surgery patients. The 66 days is the middle value a 2024 review reports for one 2010 study of simple health habits, where individuals ranged from 18 to 254 days. Expect about two months and be ready for longer.
Why do some habits take so much longer than others?
In the 2010 study, simple actions like drinking water became automatic sooner than more effortful ones like sit-ups. People also differ a great deal, and in the 2024 review the time varied from 4 to 335 days. A complex habit has more parts to become automatic.
Can you speed up habit formation?
No shortcut has been shown. The most consistent finding is to repeat the behaviour after the same cue. Keeping it simple (the 2010 study), choosing it yourself and doing it in the morning (the 2024 review) may help, though each rests on limited evidence.
How do researchers know a habit has formed?
They ask people to rate how automatic the behaviour feels, for example whether they do it without having to think about it, and track those ratings over time. In the 2010 study, the habit counted as formed when a person’s fitted curve reached 95% of its plateau.
Sources
- Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998–1009. (Online 2009.)OBSERVATIONAL, DAILY SELF-REPORT · VOLUNTEERS, ONE EATING, DRINKING OR ACTIVITY HABIT, 12 WEEKS · N = 96doi.org/10.1002/ejsp.674
- Gardner, B., Lally, P., & Wardle, J. (2012). Making health habitual: The psychology of “habit-formation” and general practice. British Journal of General Practice, 62(605), 664–666.NARRATIVE REVIEW · GUIDANCE FOR CLINICIANSdoi.org/10.3399/bjgp12X659466
- Singh, B., Murphy, A., Maher, C., & Smith, A. E. (2024). Time to form a habit: A systematic review and meta-analysis of health behaviour habit formation and its determinants. Healthcare, 12(23), 2488.SYSTEMATIC REVIEW AND META-ANALYSIS · ADULTS, HEALTH BEHAVIOURS · 20 STUDIES, N = 2,601doi.org/10.3390/healthcare12232488
- Keller, J., Kwasnicka, D., Klaiber, P., Sichert, L., Lally, P., & Fleig, L. (2021). Habit formation following routine-based versus time-based cue planning: A randomized controlled trial. British Journal of Health Psychology, 26(3), 807–824.RANDOMISED CONTROLLED TRIAL · ADULTS AGED 18–77, ONE NUTRITION HABIT, 84 DAYS · N = 192doi.org/10.1111/bjhp.12504
Sources checked .
General information, not medical advice.
How sources are chosen and checked: editorial policy.